【摘要】 通过分析比较不同配准方法的原理及优缺点,提出了一种基于图像区域特征的配准方法。算法首先对待配准图像进行自适应阈值分割,然后利用数学形态学方法进行区域轮廓提取优化,接着计算各区域的特征描述,最后以各特征向量距离最近区域的重心作为控制点集对图像进行配准。实验结果表明,该算法能够对图像进行快速准确的配准,而且具有纠正一定几何畸变的能力,是一种有效的自动配准方法。更多还原

【Abstract】 A new image registration method based on region features is proposed,through analyzing the principle,advantages and dis-advantages of the methods commonly used in image registration.Image segmentation is used firstly to extract close region regions,and then region features are calculated and feature matching is made by feature distance vector.Experimental results show that the method could accomplish image registration fast and accurately,and could correct a definite geometrical distortion.更多还原